The text fading defect is one of the most common defects in electrophotographic printers; and it dramatically affects print quality. It usually appears in a significant symbol Region of Interest (ROI), easily noticed by a user on his or her print. We can detect text fading by the density reduction for the black and white printed symbol ROI. It is difficult to detect the color text fading only by density reduction, because the depleted cartridge may only cause the color distortion without density reduction in the color printed symbol ROI. In our previous work with print quality defects analysis, the text fading detection method only works for black text fading defect detection [1]. Our new text fading method can detect the color text fading defect and predict the depleted cartridge. In this new text fading detection method, we use whole page image registration and the median threshold bitmap (MTB) matching method to align the text characters between the master and test symbol ROIs, because with the aligned text characters, it is easy to extract the difference between the master and the test text characters to detect the text fading defect. We use a support vector machine classifier to assign a rank to the overall quality of the printed page. We also use the gap statistic method with the K-means clustering algorithm to extract the different text characters’ different colors to predict the depleted cartridge.